Reply to letter by Ferrante di Ruffano et al.: Patient outcomes in randomized comparisons of diagnostic tests are still the ultimate judge.